JP2002027446A - Monitoring system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To largely reduce the transmission data amount from a plurality of cameras to an image processor for generating a composite image without lowering the quality of the image. SOLUTION: A resolution specifying unit 260 specifies a resolution needing in the case of compositing images for each region of each camera image according to a corresponding relationship between the composited image described in a mapping table 220 and each camera image. A compressing unit 120 provided in a camera 10 compresses camera image data in response to the specified resolution. The image data of the camera image compressed according to the corresponding relationship between the composite image and the camera image is transmitted to a transmission route 15.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing technique for synthesizing images taken by a plurality of cameras by processing such as deformation and integration, and more particularly, to a monitoring system for supporting driving operation of a vehicle. Belong to effective technology.

[0002]

2. Description of the Related Art In recent years, with the spread of in-vehicle display devices and the reduction in the price of video equipment such as cameras, devices that support safe driving by monitoring the periphery of a vehicle with a camera have been put into practical use and are becoming widespread. is there.

As an example of such a vehicle periphery monitoring device,
There has been proposed a system that utilizes images from a plurality of cameras installed in a vehicle, synthesizes an image as if the camera was installed directly above the vehicle, and displays the synthesized image as if the image was displayed to a driver. See Japanese Patent Application No. 10-217261).

FIG. 19 is a diagram showing a configuration example of this system. An image signal of one field or one frame is output from the plurality of cameras 401 in the camera unit 40, and the output image signal is transmitted via the transmission line 45 and stored in the buffer memory 501 of the image processing unit 50. The image synthesizing unit 503 generates a synthesized image from the image signal stored in the buffer memory 501 according to the data of the mapping table 502 that describes the relationship between the synthesized image viewed from the virtual viewpoint and the actual image of each camera 401, and displays the generated image. It is displayed on the device 60. By using the system shown in FIG. 19, the driver can accurately grasp the positional relationship between the own vehicle and the surroundings by the composite image without considering the actual camera installation position. It can be done safely.

[0005]

However, the above-described system has the following problems.

[0007] As can be seen from FIG. 19, the transmission line 45 is provided for each camera, and is disposed between each camera 401 and the image processing unit 50. Since each camera 401 is usually installed at various places around the vehicle for generating a composite image, the transmission line 45 also needs to be routed around each part of the vehicle. For this reason, complicated work is required to install this system in the vehicle, and maintenance work for troubles and the like becomes enormous.

[0007] In order to facilitate the installation and maintenance of the system, it is desirable to share the transmission lines among the cameras and reduce the number of transmission lines. However, a large transmission capacity is required to transmit all the image data of each camera, which makes it difficult to reduce the number of transmission lines. That is, in order to reduce the number of transmission lines, it is necessary to reduce the amount of transmitted image data.

The image processing section 50 also needs to store one field or one frame of image data from each camera, so that the buffer memory 501 requires a large memory capacity.

[0009] On the other hand, the camera image does not use all areas for image synthesis, and includes portions unnecessary for image synthesis. Also, the resolution required to generate a composite image differs for each part, even within the area required for image composition. Therefore, it is not always necessary to transmit all the image data of the camera image to the image processing unit as it is.

In view of the above problems, the present invention provides a monitoring system including a plurality of cameras and an image processing unit that generates a composite image from images captured by these cameras without reducing the quality of the composite image. It is an object to significantly reduce the amount of data of a transmitted camera image.

[0011]

Means for Solving the Problems In order to solve the above-mentioned problems, a solution taken by the invention of claim 1 is a surveillance system in which a camera section having a plurality of cameras and an output from the camera section are provided. The camera comprising: a transmission path for transmitting image data of a camera image; and an image processing unit that receives image data of each camera image transmitted via the transmission path and generates a composite image from the image data. The unit includes image data reduction means for reducing the amount of image data to be output to the transmission path, and the camera unit or the image processing unit, for the image data reduction means, the combination of the synthesized image and each camera image The image processing apparatus further includes a reduction mode designating unit for designating a data mode reduction mode of a camera image used for image composition according to the correspondence relationship.

According to the first aspect of the present invention, the image data reducing means provided in the camera unit is adapted to reduce the data amount of the camera image specified by the reduction mode specifying means.
The amount of image data output to the transmission path is reduced. And
The reduction mode designating means specifies a data mode of the camera image used for image synthesis in accordance with the correspondence between the composite image and the camera image. For this reason, the image data of the camera image reduced according to the mode specified according to the correspondence between the composite image and the camera image is transmitted to the transmission path. As a result, the data amount of the transmitted camera image is significantly reduced, and the quality of the synthesized image is not reduced. As a result, since the transmission capacity of the transmission path can be reduced, the transmission path can be realized by using inexpensive and small-number transmission lines or wirelessly, and the work of installation and maintenance on the vehicle is greatly simplified. . Further, the storage capacity of the buffer memory required for the image processing unit can be significantly reduced.

According to a second aspect of the present invention, in the monitoring system according to the first aspect, the image processing unit is configured to be capable of generating a plurality of types of composite images and to switch the type of the generated composite image. The reduction mode designating means switches the data amount reduction mode to be specified according to the type of the composite image generated by the image processing section.

[0014] According to the third aspect of the present invention, the first aspect is provided.
In the monitoring system, the reduction mode designating unit has a resolution designating unit for designating a resolution assumed to be necessary for generation of a composite image for each area of a camera image used for image composition, and The means compresses the image data of the camera image used for image synthesis according to the resolution specified by the resolution specifying unit.

According to the third aspect of the present invention, since the data amount of the transmitted camera image is greatly reduced by the compression according to the resolution required for the image synthesis, the transmission capacity of the transmission path can be reduced. . In addition, the occurrence of aliasing distortion can be suppressed, and the quality of the synthesized image can be improved.

According to a fourth aspect of the present invention, in the surveillance system according to the third aspect, the image data reducing means includes a DC
It is assumed that image data is compressed using T conversion.

Further, in the invention of claim 5, according to claim 1,
In the monitoring system of the above, the reduction mode designating means has an area designating unit for designating an area that is assumed to be necessary for generation of a synthesized image for a camera image used for image synthesis,
The image data reduction unit deletes image data of an area other than the area specified by the area specifying unit from the image data of the camera image used for image synthesis.

According to the fifth aspect of the present invention, the data amount of the transmitted camera image is greatly reduced by deleting the image data in the area other than the area required for the image synthesis, so that the transmission capacity of the transmission path is reduced. Need less.

According to the sixth aspect of the present invention, the first aspect is provided.
In each of the monitoring systems, each of the cameras is configured so that the reading order of the image data can be controlled from the outside, and the camera unit or the image processing unit sets the reading order of the image data of the camera used for the image synthesis, It is assumed that a reading control unit for controlling according to the data amount reduction mode specified by the reduction mode specifying unit is provided.

According to the sixth aspect of the present invention, the order of reading the image data of the camera used for image synthesis is controlled in accordance with the specified data amount reduction mode. Concentration can be prevented, and conversely, the amount of data can be dispersed over time.

[0021] In the invention of claim 7, the first aspect is provided.
The plurality of cameras in the monitoring system of (1) are installed in a vehicle and photograph the periphery of the vehicle.

[0022]

Embodiments of the present invention will be described below with reference to the drawings.

(First Embodiment) FIG. 1 is a block diagram showing a configuration of a monitoring system according to a first embodiment of the present invention. The monitoring system shown in FIG. 1 is mounted on a vehicle, and is assumed to be used for assisting a driving operation at the time of parking or the like. That is, the system generates a composite image as if it were captured by a camera installed above the vehicle, for example, by using images from a plurality of cameras installed in the vehicle, and displays it to the driver. The driver can accurately grasp the relationship between his vehicle and the surrounding situation by looking at the displayed composite image, and thus can safely perform a parking operation or the like.

In FIG. 1, a camera unit 10 has X cameras (camera 1 to camera X) 110, and each camera 110 has a compression unit 12 for compressing a camera image.
0 and the transmission adapter 130 are integrated. Image data of each camera image output from the camera unit 10 is:
The image is input to the image processing unit 20 via a transmission line 15 as a transmission path connecting the camera unit 10 and the image processing unit 20. The image processing unit 20 performs processing such as deformation and integration on the image data of each camera image, generates a composite image, and displays the composite image on the display device 30.

FIG. 2 shows an example of a camera installation position according to this embodiment. In the example of FIG. 2, six cameras 1 to 6
Are arranged around the vehicle. FIG. 3 is an example of the position of the virtual viewpoint of the composite image. The image processing unit 20 generates, for example, a composite image viewed from a virtual viewpoint as shown in FIG.

The image processing section 20 has a mapping table 220 that describes the correspondence between the synthesized image and each camera image.
And an image synthesizing unit 210 that generates a synthesized image using the mapping table 220, a decompression unit 240 that decompresses the compressed image data input via the transmission adapter 250, and temporarily stores the decompressed image data. And a buffer memory 230 for performing the operation. A reduction mode designating unit for designating, for each region of each camera image, a resolution assumed to be necessary for generation of a composite image according to the correspondence between the composite image described in the mapping table 220 and each camera image; Also, a resolution designation unit 260 is provided.

FIG. 4 is a diagram showing an example of the correspondence between the camera image and the synthesized image. The virtual viewpoint shown in FIG. 3 is obtained from the images taken by the six cameras 1 to 6 installed as shown in FIG. It is assumed that a composite image from the viewpoint of 1 is generated.
In FIG. 4A, CA1 to CA6 are areas occupied by images captured by the cameras 1 to 6 on the composite image, respectively. Further, in FIG.
A6 'is an area on each camera image corresponding to areas CA1 to CA6 on the composite image.

FIG. 4 shows the mapping table 220.
The correspondence between the composite image and each camera image as shown in FIG. That is, the mapping table 220 describes the data of the corresponding camera image for all coordinates of the composite image viewed from the virtual viewpoint. The image composition unit 210 generates a composite image viewed from a virtual viewpoint according to the data of the mapping table 220.

For example, for the point P1 on the composite image,
Since the photographing areas CA1 and CA6 of the cameras 1 and 6 are in the overlapping area OA, the area C on the image of the camera 1
Pixel data is obtained using pixel data corresponding to the point P1 in A1 'and pixel data corresponding to the point P1 in the area CA6' on the image of the camera 6 on the camera 6. Also, a point P on the composite image
2 is in the photographing area CA1 of the camera 1,
Pixel data is obtained using the pixel data corresponding to the point P2 in the area CA1 'on the image of the camera 1.

In addition, the resolution specifying section 260 determines, for each area of each camera image, each area according to the correspondence between the composite image described in the mapping table 220 and each camera image.
Find the resolution required for image composition. The data of the determined resolution is transmitted from the transmission adapter 250 to each compression unit 120 via the transmission line 15. Each compression unit 120 compresses image data of each camera image according to the transmitted resolution data. The compressed image data is transmitted from the transmission adapter 130 to the image processing unit 20 via the transmission line 15 and decompressed by the decompression unit 240 so that the buffer memory 2
30 is stored.

The mapping table 220 is shown in FIG.
The mapping table data corresponding to a plurality of virtual viewpoints as shown in (1) is stored, and the mapping table data used for image synthesis can be switched by a switching signal given from the outside. This allows
The image processing unit 20 can generate a plurality of types of composite images, and can switch the type of the generated composite image. The switching signal is given according to, for example, the state of the gears of the vehicle, the traveling speed, and the like.

When the switching signal is input, the mapping table data output from the mapping table 220 is switched, whereby the type of the generated composite image is switched. The resolution specifying unit 260 newly obtains a resolution required for image synthesis for each area of each camera image according to the new mapping table data that has been switched. The data of the determined resolution is transmitted to each compression unit 120 via the transmission line 15, and each compression unit 120 switches the content of the compression processing according to the transmitted resolution data.

<Compression Processing> Here, each compression unit 120
Performs compression of camera image data using DCT transformation.

First, a camera image (480 × 720) is divided into blocks of (8 × 8) pixels. By this division, the data of the coordinates (i, j) of the camera image is (i = 1 to 1).
480, j = 1 to 720), S1 (K, L, i ',
j ′). Here, K and L are horizontal and vertical address numbers indicating each block, and K = 1 to 6
0, L = 1 to 90. Also, i ′ = 1 to 8, j ′ =
1 to 8.

Then, each pixel data S1 (K, L,
i ′, j ′) is converted into a signal g1 (K, L, m, n) of each frequency component by the DCT transformation shown in the following equation.

[0036]

(Equation 1)

In the above equation, the component having a large value of m or n represents a high-frequency component in the horizontal or vertical direction in the block, respectively. By deleting high-frequency component data, low-resolution data can be easily generated, and a camera image can be transmitted with a small data amount. Depending on the type of the synthesized image, the required resolution may be different between the horizontal direction and the vertical direction. In this case, too, by setting the value of m or n for deleting data individually, The required resolution can be controlled independently in the vertical direction and the vertical direction.

FIG. 5 is a diagram conceptually showing the correspondence between the synthesized image and the image of the camera 1. In the figure, (a) shows a map table of the camera 1 on the composite image, (b) shows a map table of the composite image on the image of the camera 1,
The points A and A 'and the points B and B' correspond to each other. In FIG. 5A, in a portion where the grid of the map table of the camera 1 is dense (for example, point A),
Although not very high resolution is required for the image data of the camera 1, high resolution image data is required in a portion where the grid is coarse (for example, point B). That is, in FIG. 5B, high resolution image data is required in a portion where the grid of the map table of the composite image is dense (for example, point B ′ corresponding to point B), but a coarse portion (for example, point A). In the point A ′) corresponding to the above, image data of low resolution is sufficient. Further, the image data of the portion of the composite image map table where the grid is not formed is not necessary to generate the composite image.

That is, the image data of the camera 1 is divided into three types of blocks BL1, BL2, as shown in FIG.
BL3. BL1 is a block of a portion unnecessary for generating a composite image, BL2 is a block of a portion requiring high-resolution image data, and BL3 is a block of a portion requiring only low-resolution image data.

The resolution specifying section 260 calculates required resolution data Rxv (K, L) and Rxh (K, L) for each block from the data in the mapping table 220. Here, x indicates a camera number (1 ≦ x ≦ X), and subscripts h and v indicate a horizontal or vertical direction. Resolution data R
An example of a method of calculating xv (K, L) and Rxh (K, L) will be specifically described with reference to FIG.

First, it is determined from the data in the mapping table 220 whether or not there is a point corresponding to any of the coordinates of the composite image for each block of each camera image.
A block that does not include a point corresponding to the coordinates of the composite image, that is, a block BL1 is unnecessary for generation of the composite image, and thus the resolution data Rxv (K, L) and Rxh (K,
L) is “0”.

The coordinates (i, j) of the composite image (FIG. 7A)
Are present in the block when the point (i ′, j ′) corresponding to the three points (i + 1, j),
Points corresponding to (i + 1, j + 1) and (i + 1, j) are obtained. Then, an area IM1 (FIG. 7B) surrounded by these three points and the point (i ′, j ′) is determined. Then, using the vertical and horizontal sizes Lv and Lh (however, integers) of the area IM1 as shown in FIG.
The resolution data Rxv (K, L), Rxh
(K, L) are calculated respectively. However, the remainder is rounded up. Rxv (K, L) = 8 / Lv Rxh (K, L) = 8 / Lh The resolution specifying unit 260 calculates the resolution data R1v (K, L), R1h (K, L) to RXv
(K, L) and RXh (K, L) are transmitted to the compression unit 120 of each camera via the transmission line 15.

Strictly speaking, the coordinates (i,
The area occupied by the pixel j) on the camera image is the area IM1
Instead of the coordinates (i-0.5, j-0.
5), (i−0.5, j + 0.5), (i + 0.5, j)
−0.5) and (i + 0.5, j + 0.5). However, this region and the above-described region IM1 have substantially the same shape,
Also, points (i, j), (i + 1, j),
Points corresponding to (i + 1, j + 1) and (i + 1, j) can be easily obtained from the mapping table 220. Therefore, here, the resolution data Rxv (K,
L) and Rxh (K, L) are calculated from the size of the area IM1.

Each compression unit 120 outputs a signal gx (K, K, L) of each frequency for each block of each camera according to the transmitted resolution data Rxv (K, L), Rxh (K, L).
L, m, n), unnecessary components are deleted, and only necessary components are transmitted to the image processing unit 20. Specifically, a component in which the value of m is larger than Rxv (K, L),
The components larger than h (K, L) are deleted. At this time, for the block BL1, the resolution data Rxv
Since both (K, L) and Rxh (K, L) are “0”, all components are deleted.

FIG. 8 is a diagram showing an example of the internal configuration of the compression section 120. FIG. 8 shows a compression unit 120 provided for the camera 1. The compression unit 120 is provided with three 8-line memories 121a, 121b, and 121c, and can store image signals for a total of 24 (= 8.3) lines.

Resolution data R1v (K, L), R1h
When (K, L) to RXv (K, L) and RXh (K, L) are transmitted through the transmission line 15, the transmission adapter 130
Are resolution data related to the corresponding camera (in this example, resolution data R1v (K, L), R
1h (K, L)) and receives the resolution data memory 12
4 is stored.

On the other hand, from the camera 110, image signals are sent in time series according to scanning lines like television signals. Therefore, the time t during which a signal at a certain coordinate (i, j) on the screen is input is determined by the following equation. However,
Tpix is the time per pixel, and Bh is the number of horizontal blanking pixels. t = Tpix · (i · (720 + Bh) + j) The image signal sent from the camera 110 is sequentially stored in each of the eight-line memories 121a, 121b, and 121c. At this time, the image signal of the i-th line is stored in the mod (i, 24) -th line (mod (K, L) represents the remainder when K is divided by L).

When the image signals up to the eighth line are stored in the first eight-line memory 121a, the image signals after the ninth line start to be stored in the second eight-line memory 121b. On the other hand, the DSP 122 converts the (8 × 720) pixel image signal from the first eight line memory 121a into 90 8 × 8 pixel block signals S1 (1, L, i′j ′).
(L = 1 to 90, i ′ = 1 to 8, j ′ = 1 to 8), perform DCT transformation on each block as described above, and calculate g1 (1, L, m, n). . DSP
122 further uses the data R1v (1, L) and R1h (1, L) stored in the resolution data memory 124,
From g1 (1, L, m, n), a component in which m is larger than R1 v (1, L) and a component in which n is larger than R1 h (1, L) are deleted.

Therefore, the first 8-line memory 121
a stored in the image signal S1 (i, j) (i = 1 to 8,
(j = 1 to 720) at this point are the following data strings having different numbers of data for each block. g1 (K = 1, L = 1 to 90) (L = 1): d1, d2,..., dmn (L = 2): d1, d2,..., dmn '(L = 90): d1, d2..., dmn ″ (mn = R1v (1,1) · R1h (1,1) mn ′ = R1v (1, 2) · R1h (1, 2) mn ″ = R1v (1, 90) · R1h ( 1, 90)) That is, the image signal stored in the first eight-line memory 121a is (R1v (1, L) ·
It is converted into a data string having R1h (1, L) data. This data sequence is temporarily stored in the DCT data buffer memory 123, and the transmission adapter 130 transmits the data sequence stored in the DCT data buffer memory 123 to the image processing unit 20 via the transmission line 15.

The restoration unit 240 performs an inverse DCT transform on the transmitted data sequence and performs image signal S1 '(1, L, i', j ').
Is restored and stored in the buffer memory 230. The restored signal S1 ′ (1, L, i ′, j ′) has the high-frequency component removed during the DCT transform, so that the original signal S1 ′ (1, L, i ′, j ′) has been deleted.
(1, L, i ', j') corresponds to a signal that has been subjected to LPF processing.

Next, when the image signals up to the 16th line are stored in the second 8-line memory 121b, the image signals after the 17th line are started to be stored in the third 8-line memory 121c. On the other hand, the DSP 122 converts the (8 × 720) pixel image signal from the second 8-line memory 121b into 90 8 × 8 pixel block signals S1 (2, L, i ′).
j ′) (L = 1 to 90, i ′ = 1 to 8, j ′ = 1 to 8)
, And for each block, as described above, D
The CT conversion is performed to calculate g1 (2, L, m, n).
The DSP 122 further uses the data R1v (2, L), R1h (2, L) stored in the resolution data memory 124, and obtains g from R1v (2, L, m, n).
The component larger than L) and the component where n is larger than R1h (2, L) are deleted.

Therefore, the second 8-line memory 121
At this point, the image signal stored in b becomes a data string having a different number of data for each of the following blocks. g1 (K = 2, L = 1 to 90) (L = 1): d1, d2, ..., dmn (L = 2): d1, d2, ..., dmn '(L = 90): d1, d2 ", mn" = R1v (2,1) .R1h (2,1) mn '= R1v (2,2) R1h (2,2) mn "= R1v (2, 90) .R1h ( 2, 90)) That is, the image signal stored in the second 8-line memory 121b is (R1v (2, L) ·
It is converted into a data string having R1h (2, L)) data. This data string is transmitted to the image processing unit 20 via the transmission line 15 in the same manner as described above, and the image signal S1 ′ (2, L, i ′, j ′) is restored by the restoration unit 240.
Is restored.

The same operation is repeated until K = 60, whereby the image signal S1 '(K, L, i',
j ′) is restored and stored in the buffer memory 230.

Similarly, the compressed image data is transmitted from the compression unit 120 of the other camera 110 to the decompression unit 240, and the image signal Sx ′ (K, L, i ′, j ′) (x =
2 to X) are restored and stored in the buffer memory 230. From each camera image data stored in the buffer memory 230, a composite image is generated as in the conventional case.

As described above, according to the present embodiment, only the data of the required resolution in the image signal of each camera is compressed and transmitted according to the correspondence between the composite image described in the mapping table 220 and each camera image. Therefore, the transmission capacity of the transmission line 15 can be smaller than in the related art. For this reason, a device capable of more stable transmission can be realized, and a cheaper transmission line can be used.

That is, as shown in FIG.
Can be easily integrated into one, and the installation and maintenance of the transmission line 15 in the vehicle can be greatly simplified. Alternatively, FIG.
As shown in FIG.
Can be flexibly handled, for example, by dividing
Installation and maintenance on the vehicle can be further simplified.

Note that, instead of the transmission line 15, a transmission path may be configured wirelessly. Even in this case, the effect of requiring a small transmission capacity, which is obtained by the present embodiment, is great, and it is possible to realize a transmission path using less expensive components. In addition, installation and maintenance work on the vehicle can be further significantly reduced as compared with the case of wired communication.

For portions unnecessary for generating a composite image as shown in a block BL1 in FIG. 6B, data Rxv (K, L) and Rxh (K,
Since L) is (0,0), the number of data in the block in this portion is zero. Therefore, the amount of data transmitted via the transmission line 15 can be significantly reduced.

In the image processing section 20, only the data of the required resolution among the camera images is stored in the buffer memory 23.
Since it is stored in 0, the memory capacity can be significantly reduced as compared with the related art.

In the compression unit 120, DCT conversion and transmission are performed immediately after the camera image signal is stored in each of the eight line memories 121a, 121b, and 121c, so that the signal delay due to compression transmission is minimized. Can be suppressed.

The compression operation of each camera image can be switched according to the type of the composite image generated by the image processing section 20. In this case, each time the mapping table used for image synthesis is switched, the resolution specifying unit 26
0 is the resolution data Rxv (K, L), Rxh (K,
L) may be calculated, and this data may be transmitted to the compression unit 120 of each camera via the transmission line 15.

Alternatively, resolution data corresponding to each mapping table may be stored in advance in the resolution specifying section 260 by a ROM or the like. In this case, the switching signal may be input to the resolution specifying unit 260, and the resolution data may be switched by the resolution specifying unit 260 together with the switching of the mapping table. This eliminates the need to calculate the resolution data each time the mapping table is switched.

Further, a memory is provided in the compression unit 120,
In this memory, resolution data corresponding to each mapping table may be stored in advance. In this case,
The image processing unit 20 switches the mapping table
Only the ID representing the mapping table is assigned to each compression unit 12
It only needs to be transmitted to 0. Alternatively, instead of transmitting the ID representing the mapping table from the image processing unit 20,
The switching signal may also be input to each of the compression units 120 of the camera unit 10 so that each of the compression units 120 executes the switching of the resolution data together with the switching of the mapping table.

In this embodiment, the DCT is used to compress the image data. However, the data after the DCT may be further quantized like JPEG. In addition to the DCT transform, a similar effect can be obtained by a wavelet transform, a Fourier transform, or the like.

<Effect of Eliminating Aliasing Distortion> Further, according to the present embodiment, it is possible to easily avoid the deterioration of the image quality of the synthesized image due to the aliasing distortion.

For example, referring to FIG.
The area IM1 on the camera image corresponding to (i, j) is
Multiple pixels larger than one pixel of the camera image
Includes To obtain an optically accurate composite image
Is a weighted average of data of a plurality of pixels included in the area IM1.
Value as pixel data corresponding to coordinates (i, j).
I have to. That is, as shown in the following equation,
It is necessary to determine the pixel data S (i, j). S (i, j) = Σγ_p・ (S1 (i_p, J_p)) Where S1 (i_p, J_p) Is camera pixel data, p
Is the camera pixel number, γ _pIs the ratio included in the area IM1
Is a factor in which However, in this case,
Buffer the signals of multiple pixels to obtain a signal of one pixel
It is necessary to read from the memory 230 and perform the calculation. This
This is because the amount of calculation processing required by the image
Is not practical.

Therefore, the nearest method has been conventionally used as an approximate omission of the above calculation. The nearest method is an area IM1 corresponding to the coordinates (i, j).
Is used only the signal of one pixel closest to the center of.

However, when the nearest method is used, a unique aliasing distortion occurs when a signal having a high sampling frequency is sub-sampled at a low sampling frequency without cutting off high-frequency components. That is, when a simple image synthesis is performed by the nearest method using the high-resolution image signal as it is, there is a possibility that aliasing distortion occurs in a portion where a low resolution is required and the image quality of the synthesized image is deteriorated. is there.

In order to prevent such aliasing distortion, for example, as shown in FIG. 10, an image signal of each camera is subjected to processing by a low-pass filter (LPF) 51 in advance.
A method of cutting unnecessary high-frequency components can be considered.

In this case, however, since the resolution required in the camera image differs for each part, it is necessary to use an LPF whose frequency characteristic can be set adaptively for each part of the image. Therefore, there is a problem that the circuit configuration becomes considerably complicated. When an LPF having a uniform frequency characteristic is used, the circuit configuration is simplified, but high-frequency components are cut off even in a portion requiring a high resolution. I can't.

On the other hand, according to the present embodiment, for the low-resolution image portion where aliasing distortion is likely to occur, the high-frequency component is already removed by the compression section 120 with characteristics corresponding to the size of the region IM1. Therefore, aliasing distortion is less likely to occur. That is, the camera image signal restored by the restoration unit 240 corresponds to a signal obtained by adaptively performing LPF processing on the original signal in accordance with a required resolution. Therefore, the occurrence of aliasing distortion can be suppressed without providing an LPF having a complicated configuration, and as a result, the image quality of a composite image can be significantly improved.

Although the embodiment using the nearest method has been described, other methods can be applied. For example, it is obviously applicable to a bilinear method (a method of performing linear interpolation using signals of four pixels surrounding the center of the area IM1 corresponding to the coordinates (i, j)).

<Another Example of Compression Processing> In the above description, the resolution specifying section 260 outputs the resolution data Rxv (K,
Although L) and Rxh (K, L) are calculated from the horizontal and vertical sizes Lh and Lv of the area IM1 shown in FIG. 7, they may be calculated by other methods.

FIG. 11A shows an area IM1 occupied by one pixel of the composite image on the camera image, similarly to FIG. 7B. Here, as shown in FIG.
8) The area IM1 is placed at the center of the block, and for each pixel, a coefficient γ _p considering the ratio included in the area IM1 is obtained. The coefficient γ _p corresponds to a coefficient used for weighting each pixel when a weighted average value of a plurality of pixels of the camera image is used to obtain an optically accurate composite image.

The coefficients r (i ′, j ′) are developed into 8 × 8 block coefficients with the center of the area IM1 being (i ′, j ′) = (1, 1). And, like the image signal, the DCT
The conversion is performed to obtain a conversion coefficient h (K, L, m, n). The conversion coefficient h (K, L, m, n) is calculated by the LPF indicated by the coefficient γ _p
The characteristic is represented by a DCT transform coefficient, and the high frequency component is reduced.

Therefore, the upper limits of m and n whose conversion coefficients h (k, L, m, n) are equal to or greater than a predetermined threshold value are defined as the values of the resolution data Rxv (K, L) and Rxh (K, L). Shall be defined as The resolution specifying unit 260 outputs the resolution data Rx
v (K, L) and Rxh (K, L)
A conversion coefficient h (K, L, m,
n) is also transmitted to the compression unit 120 at the same time.

The compression section 120 converts the DCT data gx (K, L, m, n) of the image signal into resolution data Rx
High frequency components are deleted using xv (K, L) and Rxh (K, L). Then, the remaining DCT transformed data gx
(K, L, m, n) is converted to a conversion coefficient hx (K, L, m, n)
And sends the result to the restoration unit 240.

In the case of this method, the restoration unit 240 performs the inverse DCT
The signal restored by the processing is obtained by approximately performing LPF processing on the original signal using the coefficient γ _p . Therefore, an optically more accurate composite image can be obtained as compared with the above-described embodiment.

The resolution data Rxv (K, L) and Rxh (K, L) can be determined without using the area IM1. For example, the resolution data Rxv (K, L), Rxh
(K, L) is set to “0”, and the resolution data Rxv (K, L), Rxv
h (K, L) may be set to a uniform value other than “0”.

(Second Embodiment) FIG. 12 shows a second embodiment of the present invention.
It is a block diagram showing the composition of the monitoring system concerning an embodiment. The monitoring system shown in FIG. 12 is basically the same as the configuration according to the first embodiment shown in FIG. FIG.
The same components as those in FIG. 1 are denoted by the same reference numerals as those in FIG. 1, and detailed description thereof will be omitted.

12 differs from FIG. 1 in that the image processing unit 20A generates a read control signal for each camera 110A in accordance with the output of the resolution designation unit 260, and the read control unit 270 and each camera 110A And a synchronization signal generation unit 280 for generating the synchronization signal. The read control signal and the synchronization signal output from the read control unit 270 and the synchronization signal generation unit 280 are transmitted to the transmission adapter 250.
Is transmitted to the camera unit 10A via the transmission line 15.
The transmitted read control signal and synchronization signal are transmitted to the compression unit 120A and the camera 110A via the transmission adapter 130.
Sent to

Each camera 110A captures a frame image at the same timing in accordance with the transmitted synchronization signal.
Here, it is assumed that each camera 110A captures image data of (720 × 480) pixels every 1/60 second.

FIG. 13 shows the camera 110A and the compression unit 12
It is a figure showing an example of an internal configuration of 0A. As shown in FIG. 13, the camera 110A mainly includes an image sensor (here, CC
D) 111, optical system lens 112, readout / synchronization control unit 1
13 and an AD conversion processing unit 114 that performs signal processing such as color separation after AD-converting the output of the image sensor 111.

The image pickup device 111 includes an image pickup surface for one screen and one image pickup surface.
It has a storage system for the screen. First, the optical system lens 11
2, a light intensity signal is obtained on the imaging surface, and the phototransistor of each pixel converts the light intensity signal into a charge signal. After the converted charge signal is stored for one frame period (here, 1/60 second), it is transmitted to the storage system as an image signal.

In a normal CCD, the image signal stored in the storage system is read from the upper left end of the screen according to the scanning lines. Meanwhile, on the imaging surface, a new light intensity signal is converted into a charge signal and stored as an image signal of the next frame. However, in the present embodiment, unlike this, the image sensor 111 is configured to be able to read the image signal from the upper left end or the lower left end of the screen. In which order the image signals are read out,
It is assumed that it is controlled by the read / synchronization control unit 113.

In this embodiment, the image processing unit 20A controls the reading order of the image data of each camera 110A according to the correspondence between the synthesized image and each camera image.

FIG. 14 is a diagram showing an example of the reading order of image signals from each camera 110A. In FIG. 14, areas CA1 ′ to CA6 ′ of each camera image are shown in FIG.
This is the same as that shown in (b), and is an area necessary for generating a composite image. As can be seen from FIG. 14, in each camera image, the area required for image composition occupies the lower part of the screen, and the upper part of the screen is out of the area required for image composition.

Here, the read control unit 270 reads the image signals from the cameras 1, 3 and 5 from the upper left corner of the screen as indicated by the arrows in FIG.
For No. 6, a read control signal is generated so that an image signal is read from the lower left end of the screen. The read / synchronization control unit 113 of each camera 110A controls the image sensor 111 according to the transmitted read control signal. The output of the image sensor 111 is output to the A
After the D conversion, the signal is subjected to signal processing such as color separation, and is sequentially sent to the 8-line memory 121 of the compression unit 120A as an image signal.

For cameras 1, 3, and 5, as in the first embodiment, the signal at a certain coordinate (i, j) on the screen is
It is sent at time t in the following equation. t = Tpix · (i · (720 + Bh) + j) On the other hand, since the cameras 2, 4, and 6 are scanned from the lower left in the screen, certain coordinates (i, j) in the screen are different from the first embodiment. ) Is sent at time t in the following equation. t = Tpix · ((480-i) · (720 + Bh) +
j) where Tpix is the time per pixel and Bh is the number of horizontal blanking pixels.

Therefore, as for the camera 1, the first image signal is converted into a data string having a different number of data for each block as described below and transmitted, as in the first embodiment, whereas g1 ( K = 1, L = 1 to 90) (L = 1): d1, d2,..., Dmn (L = 2): d1, d2,..., Dmn '(L = 90): d1, d2 , Dmn "(mn = R1v (1,1) .R1h (1,1) mn '= R1v (1,2) R1h (1,2) mn" = R1v (1,90) .R1h (1, 90)) With respect to the camera 2, the first image signal is converted into a data string having a different number of data for each block and transmitted as follows. g2 (K = 60, L = 1 to 90) (L = 1): d1, d2, ..., dmn (L = 2): d1, d2, ..., dmn '(L = 90): d1, d2..., dmn ″ (mn = R2v (60,1) · R2h (60,1) mn ′ = R2v (60,2) · R2h (60,2) mn ″ = R2v (60,90) · R2h ( 60, 90))

Here, the amount of transmission data on the transmission line 15 when the compressed image signal is transmitted from each camera 110A will be considered.

As for the image signals of the first eight lines of the screen, mn to mn ″ are almost 0 from the cameras 1, 3, and 5, and only data of almost the same data amount is transmitted. A considerable amount of data required for image synthesis is transmitted from 2, 4, and 6. Conversely, the image signals of the last eight lines of the screen are transmitted by the cameras 2 and 4.
From 4,6, mn to mn "become almost 0, and only data having almost no data amount is transmitted. On the other hand, cameras 1, 3, 5 transmit a considerable amount of data necessary for image synthesis. Transmitted.

As described above, by reversing the reading order of the image signals from the cameras 1, 3, and 5 and the cameras 2, 4, and 6, the amount of transmission data on the transmission line 15 can be dispersed in time. it can. This makes it possible to use a transmission line 15 having a smaller transmission capacity.

Note that, even in the case of the first embodiment, it is possible to temporally disperse the amount of transmission data on the transmission line 15. For example, if the capacity of the DCT data buffer memory 123 of the compression unit 120 is made sufficiently large and data transmission is controlled on the output side of the DCT data buffer memory 123, the amount of transmitted data can be dispersed in time. .

However, in this case, it is necessary to temporarily store a large amount of data in the DCT data buffer memory 123, so that the delay time in data transmission from the camera 110 to the image processing unit 20 becomes long. Since the driver drives the vehicle while watching the composite image, it is preferable that such a delay time is short in consideration of the response of the composite image output. In this regard, the present embodiment is more effective.

As described above, according to the present embodiment, the reading order of the image signals of each camera is controlled in accordance with the correspondence between the synthesized image and each camera image, so that the response of the synthesized image output is not reduced. In addition, the amount of transmission data can be dispersed over time.

The control of the reading order of the image signals shown in FIG. 14 is merely an example.
For cameras 2 and 3 from the upper left corner of the screen,
The control may be performed such that the image signal is read from the lower left end of the screen. Also, if the correspondence between the composite image and the camera image changes, the reading order of the image signals may be changed accordingly. Further, here, the image sensor 111 is configured to be able to realize two kinds of reading orders. However, if the image sensor 111 is configured to be able to realize three or more kinds of reading orders, the combined image and each camera image can be read. An appropriate reading order may be selected according to the correspondence.

Here, the CCD is used as the image pickup device, but a CMOS device may be used instead. While a camera using a CCD element outputs signals of the entire screen according to scanning lines, a CMOS
A camera using an element can output only a signal of a part of a screen (for example, a rectangular area). For this reason, the reading order of the image signal of each camera can be finely controlled, so that the amount of transmission data can be more efficiently distributed over time.

(Third Embodiment) FIG. 15 shows a third embodiment of the present invention.
It is a block diagram showing the composition of the monitoring system concerning an embodiment. The monitoring system shown in FIG. 15 is basically the same as the configuration according to the first embodiment shown in FIG.

The present embodiment is different from the first embodiment in that, instead of compressing the image data in order to reduce the image data of each camera image, instead of compressing the image data, an area which is assumed to be unnecessary for generating a composite image is used. The point is that the image data is deleted. That is, the image processing unit 20B includes, as a reduction mode designating unit, a region designating unit 290 that designates a region assumed to be necessary for generating a composite image for each camera image, instead of the resolution designating unit 260. The selecting unit 1 deletes the image data of an area other than the area specified by the area specifying unit 290, instead of the compression unit 120.
40. Further, since no image compression is performed, the restoration section 240 is omitted from the image processing section 20B.

The area designating section 290 designates an area required for image synthesis in each camera image in accordance with the correspondence between the synthesized image converted into data in the mapping table 220 and each camera image. Information on the designated area is transmitted from the transmission adapter 250 to each selection unit 140 of the camera unit 10B via the transmission line 15.

The selecting section 140 outputs only data of a necessary area in each camera image according to the transmitted area information. The output image data is transmitted from the transmission adapter 130 to the image processing unit 20B via the transmission line 15.
In the image processing unit 20B, the transmitted image data is stored in the buffer memory 230, and the image composition unit 210 generates a composite image from the image data stored in the buffer memory 230 according to the data in the mapping table 220.

FIG. 16 is a diagram conceptually showing the relationship between the image of the camera 1 and the area ANE required for image synthesis. That is, from the position of the map table of the composite image on the image of the camera 1 shown in FIG. 16A, an area ANE necessary for generating the composite image can be obtained as shown in FIG. 16B. The area specifying unit 290 obtains a rectangular area AN1 including the area ANE, for example, as shown in FIG.
As information indicating the area AN1, the coordinates (I
s, Js) and the lower right coordinates (Ie, Je) are output. The area specifying unit 290 similarly determines a rectangular area including an area necessary for generating a composite image for the other cameras, and outputs the coordinates of the upper left and lower right corners.

FIG. 18 is a diagram showing an example of the internal configuration of the selection section 140. The operation of the selection unit 140 will be described with reference to FIG.

The image signal output from the camera 110 is
The data is sequentially written to the three line memories 141. on the other hand,
The transmission adapter 130 reads only the area data corresponding to its own camera from the area data sent via the transmission line 15 and stores the area data in the area data memory 144.

When the image data is written into a certain line memory, the DSP 142 reads out the image data from the line memory, selects only the image data included in the rectangular area stored in the area data memory 144 from the image data, and The data is stored in the buffer memory 143. The image data stored in the data buffer memory 143 is added with a header indicating a camera number, a line number, and the number of data as follows, and transmitted from the transmission adapter 130 to the image processing unit 20B via the transmission line 15. (Camera number 1) (line number 1) (data number M1): d1,..., DM1 (camera number 1) (line number 2) (data number M2): d1,..., DM2 (camera number 1) (Line number 480) (number of data M480): d1,..., DM480

The data number Mi of the line i is as follows. When Is ≦ i ≦ Ie, Mi = Je−Js + 1. Otherwise, Mi = 0. That is, the amount of data transmitted through the transmission line is significantly reduced as compared with the conventional example in which all image data is transmitted. .

In this case, the area designating section 290 determines a rectangular area AN1 including an area ANE necessary for generating a composite image, and outputs coordinate information indicating the rectangular area AN1 to each of the selecting sections 1.
40, but the present invention is not limited to this. For example, as shown in FIG. 17B, the area ANE necessary for image synthesis is set to “1”, and the unnecessary area AU
N may be binarized as "0", and run-length data scanned in accordance with a horizontal scanning line may be transmitted. In this case, each selection unit 140 transmits only the image data of the area ANE among the camera images to the image processing unit 20B according to the transmitted run-length data.

Further, similarly to the first embodiment, the operation of selecting each camera image can be switched according to the type of the composite image generated by the image processing section 20B. In this case, each time the mapping table used for the image synthesis is switched, the area specifying unit 260 calculates the area data indicating the area required for the image synthesis, and transmits this data via the transmission line 15 to the selection unit 140 of each camera. May be configured to be transmitted.

Alternatively, area data corresponding to each mapping table may be stored in the area specifying section 290 in advance by a ROM or the like. In this case, the switching signal is also input to the resolution specifying unit 290, and the switching of the area table is performed together with the switching of the mapping table.
90. This eliminates the need for processing for obtaining area data each time the mapping table is switched.

Further, a memory is provided in the selection unit 140,
In this memory, area data corresponding to each mapping table may be stored in advance. In this case, each time the mapping table is switched, the image processing unit 20B
Only the ID representing the mapping table is assigned to each selection unit 14
It only needs to be transmitted to 0. Alternatively, instead of transmitting the ID representing the mapping table from the image processing unit 20B, the switching signal is also input to each of the selection units 140 of the camera unit 10B, and the switching of the area data is performed together with the switching of the mapping table. Should be executed.

Note that the second embodiment may be realized in combination with the third embodiment. That is, the read control unit 270 according to the second embodiment is added to the image processing unit 20B shown in FIG.
And the synchronization signal generation unit 280 may be provided to control the order of reading camera images in the same manner.

In the first to third embodiments, the resolution specifying section 260 or the area specifying section 290 is replaced with the image processing section 20,
20A and 20B, but instead of this,
The same means as the resolution specifying unit 260 is used for the camera units 10 and 10.
A, or a unit equivalent to the area designation unit 290 may be provided in the camera unit 10B.

In this case, the image processing units 20, 20A, 20
B may transmit the data of the mapping table 220 from the transmission adapter 250 to the camera units 10, 10A, and 10B via the transmission line 15 every time the type of the synthesized image is switched. Alternatively, a memory is provided in each of the camera units 10, 10A, and 10B, and if the resolution data or the region data corresponding to each mapping table is stored in advance in this memory, the image processing units 20, 20A, and 20B can output the synthesized image. Each time the type is switched, only the ID representing the mapping table to be used need be transmitted to the camera units 10, 10A, and 10B. At this time, instead of transmitting the ID representing the mapping table from the image processing units 20, 20A, 20B, the switching signal is transmitted from the camera units 10, 10A, 10B.
To switch the resolution data or the area data.

Further, the read control unit 270 and the synchronization signal generation unit 280 according to the second embodiment are
Like the 60 and the area designating section 290, it may be provided in the camera section 10A.

In the first to third embodiments, the camera images of all cameras are used for image synthesis. However, only some of the cameras may be used for image synthesis. In such a case, the specification of the data amount reduction mode such as the resolution specification and the area specification may be performed only for some of the cameras used for image synthesis.

In the first to third embodiments, the compression unit 120 and the selection unit 140 as image data reduction means are
Although provided for all cameras, a configuration provided for only some cameras may be employed. That is, a camera having the compression unit 120 and the selection unit 140,
A camera that does not have the compression unit 120 or the selection unit 140 may be provided together. In this case, the specification of the data amount reduction mode such as the resolution specification and the area specification is performed by the compression unit 12.
This may be performed only for a camera having 0 and the selection unit 140 and used for image synthesis.

In the above description, the monitoring system according to the present invention is applied to a vehicle. However, the monitoring system can be similarly applied to a moving object other than the vehicle, for example, an airplane or a ship. In addition, a camera may be installed in a monitoring target other than a mobile object, for example, a store, a residence, a show room, and the like. Further, the installation positions and the number of cameras are not limited to those shown here.

Further, all or a part of the functions of the monitoring system according to the present invention may be realized using dedicated hardware or may be realized by software. It is also possible to use a recording medium or a transmission medium storing a program for causing a computer to execute all or a part of the functions of the image processing apparatus according to the present invention.

[0120]

As described above, according to the present invention, the number of transmission paths between a plurality of cameras and an image processing unit is reduced in accordance with a designated mode according to the correspondence between a composite image and each camera image. The image data of the camera image is transmitted.
Therefore, the data amount of the transmitted camera image can be significantly reduced without lowering the quality of the synthesized image. This makes it possible to realize a transmission path using inexpensive and small-number transmission lines or wirelessly, greatly simplifying installation and maintenance work on the vehicle. Further, the storage capacity of the buffer memory required for the image processing unit is greatly reduced. The reduction of the image data is realized by compression according to the resolution required for the image synthesis or by deleting the image data other than the area required for the image synthesis.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a monitoring system according to a first embodiment of the present invention.

FIG. 2 is a diagram illustrating an example of a camera installation position.

FIG. 3 is an example of a position of a virtual viewpoint of a composite image.

FIG. 4 is a diagram illustrating an example of a correspondence relationship between a camera image and a composite image.

FIG. 5 is a diagram conceptually showing a correspondence relationship between a composite image and a camera image.

FIG. 6 is a diagram illustrating an example of classification based on the resolution of a camera image area.

FIG. 7 is a diagram illustrating an example of a calculation method of resolution data.

FIG. 8 is a diagram illustrating an example of an internal configuration of a compression unit according to the first embodiment of the present invention.

FIG. 9 is a diagram showing another example of a camera installation position.

FIG. 10 is a diagram illustrating a configuration including an image processing unit having an LPF as a comparative example.

FIG. 11 is a diagram for explaining another example of a method for calculating resolution data.

FIG. 12 is a block diagram illustrating a configuration of a monitoring system according to a second embodiment of the present invention.

FIG. 13 is a diagram illustrating an example of an internal configuration of a camera and a compression unit according to a second embodiment of the present invention.

FIG. 14 is a diagram showing an example of the reading order of image signals from each camera.

FIG. 15 is a block diagram illustrating a configuration of a monitoring system according to a third embodiment of the present invention.

FIG. 16 is a diagram showing an area required for image synthesis.

FIG. 17 is a diagram illustrating a method of designating an area required for image composition.

FIG. 18 is a diagram illustrating an example of an internal configuration of a selection unit according to the third embodiment of the present invention.

FIG. 19 is a diagram showing a configuration of a conventional monitoring system.

[Explanation of symbols]

 10, 10A, 10B camera unit 15 transmission path 20, 20A, 20B image processing unit 110, 110A camera 120, 120A compression unit (image data reduction unit) 140 selection unit (image data reduction unit) 260 resolution specification unit (reduction mode specification) Means) 270 read control unit 290 area designation unit (reduction mode designation means)

 ──────────────────────────────────────────────────の Continuing on the front page (72) Inventor Masamichi Nakagawa 1006 Oaza Kadoma, Kadoma City, Osaka Prefecture Inside Matsushita Electric Industrial Co., Ltd. (72) Inventor Atsushi Morimura 1006 Kadoma Kadoma, Kadoma City, Osaka Prefecture F-term in Matsushita Electric Industrial Co., Ltd.

Claims (7)

[Claims] 1. A camera unit having a plurality of cameras, a transmission path for transmitting image data of a camera image output from the camera unit, and image data of a camera image transmitted via the transmission path as inputs. An image processing unit that generates a composite image from the image data; the camera unit includes an image data reduction unit configured to reduce an amount of image data output to the transmission path; The processing unit is characterized in that the image data reduction unit includes a reduction mode designating unit that specifies a data amount reduction mode of a camera image used for image synthesis according to a correspondence relationship between the composite image and the camera image. And monitoring system. 2. The monitoring system according to claim 1, wherein the image processing unit is configured to be capable of generating a plurality of types of composite images and to be able to switch the type of the generated composite image. A monitoring system, characterized in that the mode designating means switches a data amount reduction mode to be specified according to the type of a composite image generated by the image processing unit. 3. The surveillance system according to claim 1, wherein the reduction mode designating means designates, for each area of a camera image used for image composition, a resolution assumed to be necessary for generating a composite image. A monitoring unit, wherein the image data reducing unit compresses image data of a camera image used for image synthesis according to a resolution specified by the resolution specifying unit. 4. The monitoring system according to claim 3, wherein said image data reduction means compresses image data using DCT transform. 5. The surveillance system according to claim 1, wherein the reduction mode designating unit has an area designating unit for designating, for a camera image used for image composition, an area assumed to be necessary for generating a composite image. A monitoring system, wherein the image data reduction unit deletes image data of an area other than the area specified by the area specifying unit from image data of a camera image used for image synthesis. 6. The surveillance system according to claim 1, wherein each of the cameras has a controllable read order of image data, and the camera section or the image processing section has an image of a camera used for image synthesis. A monitoring system comprising: a read control unit that controls a data reading order according to a data amount reduction mode specified by the reduction mode specifying unit. 7. The surveillance system according to claim 1, wherein the plurality of cameras are installed in a vehicle and photograph an area around the vehicle.